Unattended fresh food e-commerce cold chain system for commercial tenants

Abstract

The invention provides an unattended fresh food e-commerce cold chain system for merchants, and belongs to the technical field of fresh food cold chains, the unattended fresh food e-commerce cold chain system comprises a production system, a centralized distribution center system, a distribution system and a central server, the sorted products are packaged into a modular split charging box, and a first sensor group is arranged in the modular split charging box and comprises a three-axis MEMS acceleration sensor, a metal oxide semiconductor gas sensor array, an optical fiber smell sensor and a communication module; and the centralized distribution center system comprises a central distribution refrigeration box which is configured to receive and fix a plurality of modular sub-packaging boxes. According to the invention, the whole process forms a closed loop through the real-time monitoring of the sensor, the intelligent decision making of the central server and the cooperative execution of all systems, and the intelligent, safe and efficient management of the whole chain from production to distribution is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of fresh produce cold chain, and in particular to an unmanned fresh produce e-commerce cold chain system for merchants. Background Art

[0002] The fresh produce cold chain refers to the process of transporting fresh produce from its production site to its consumption site, maintaining a low temperature environment to ensure its freshness, quality, and safety. Cold chain logistics involves all storage and transportation links, from fresh produce supply sites such as farms, fishing vessels, or ranches to processing plants, logistics center cold storage, distributor cold storage, retail stores, and ultimately consumers. The traditional fresh produce cold chain has the following core pain points:

[0003] 1. Insufficient monitoring capabilities: Relying solely on a single temperature parameter for monitoring, it is unable to perceive in real time changes in fresh produce quality caused by collisions, gas composition (such as ethylene), etc., resulting in safety blind spots.

[0004] 2. Weak security protection: Mechanical locks are easily damaged by physical means, and traditional electronic locks are subject to the risk of side-channel attacks, with high risks of cargo theft or data tampering.

[0005] 3. Low delivery reliability: Relying on a single GPS navigation system, positioning deviations or control failures are prone to occur in complex weather or road conditions, making it difficult to guarantee transportation safety and punctuality.

[0006] 4. Data management flaws: Centralized databases are vulnerable to attacks or internal tampering, supply chain data is opaque, consumers and merchants lack trust, and quality dispute resolution is inefficient.

[0007] 5. Inefficient resource utilization: Fixed partitions waste transportation space, passive cushioning materials (such as foam) have limited shock absorption effects, and the physical damage rate of fresh produce is high. Traditional route planning relies on static data, resulting in prominent delivery delays and energy waste.

[0008] 6. Environmental compliance risks: The extensive use of non-degradable plastic packaging does not comply with international plastic ban policies.

[0009] Therefore, an unmanned fresh food e-commerce cold chain system for merchants is proposed. Summary of the Invention

[0010] The present invention aims to solve the problems raised in the background technology and provides an unmanned fresh food e-commerce cold chain system for merchants.

[0011] The specific technical solutions are as follows:

[0012] An unmanned fresh food e-commerce cold chain system for merchants, including:

[0013] A production system configured to standardize fresh produce using automatic sorting equipment and package the sorted produce into modular packaging boxes, wherein the modular packaging boxes have a built-in first sensor group including: a three-axis MEMS accelerometer, a metal oxide semiconductor gas sensor array, a fiber optic odor sensor, and a communication module;

[0014] The distribution center system includes a central distribution refrigerated container configured to receive and secure a plurality of modular packaging boxes, wherein the refrigerated container is equipped with: a second sensor group, a multi-mode wireless communication module, and an electromagnetic locking system, wherein the second sensor group includes a quartz crystal temperature sensor, an optical fiber strain sensor, and a quantum encryption lock;

[0015] The distribution system includes an unmanned cold chain vehicle loaded with central distribution refrigerated containers, and the cold chain vehicle is equipped with a reinforcement learning path planning module and a multi-sensor fusion autonomous driving unit;

[0016] The central server integrates multi-source heterogeneous data through a federated learning framework, uses the spatiotemporal graph convolutional network (ST-GCN) to generate dynamic delivery routes, and triggers a three-level early warning mechanism based on a damage prediction model to reduce the fresh food loss rate.

[0017] As a preferred solution of the present invention, the buffer structure of the modular packaging box includes:

[0018] Shape memory alloy skeleton;

[0019] Aerogel filling layer;

[0020] When the triaxial MEMS accelerometer detects a transient impact >5g, it expands within 200ms to form a honeycomb shock-absorbing structure through the shape memory effect to improve shock absorption efficiency.

[0021] As a preferred solution of the present invention, the odor sensor adopts a dynamic sniffing algorithm to identify excessive ethylene concentration through the XGBoost model, and warns of the risk of fruit and vegetable corruption 24 hours in advance to improve the prediction accuracy.

[0022] As a preferred embodiment of the present invention, the adjustable partition plate of the central distribution refrigerated box is configured as follows:

[0023] A photoelectric sensor is installed at the end of the slide rail;

[0024] Dynamically optimize the packaging box layout based on reinforcement learning algorithms to improve space utilization.

[0025] As a preferred solution of the present invention, the automatic driving unit of the unmanned cold chain vehicle includes:

[0026] Redundant braking system;

[0027] Multimodal sensor data fusion module.

[0028] As a preferred solution of the present invention, the three-level early warning mechanism includes:

[0029] Level 1 warning: When the temperature fluctuation exceeds ±0.5°C, a suggestion to adjust the receiving time will be pushed to the merchant terminal;

[0030] Level 2 warning: When the collision intensity is detected to be greater than 3g or the ethylene concentration is greater than 100ppb, the spare sub-packaging box allocation process is started;

[0031] Level 3 warning: When encountering illegal unlocking, the following will be executed simultaneously:

[0032] Release acid electrolyte into the dispensing box;

[0033] Adjust the direction of the air outlet of the cold chain vehicle's air conditioner to form a local low-temperature barrier.

[0034] As a preferred solution of the present invention, the reinforcement learning algorithm of the path planning module includes:

[0035] State space definition: traffic flow, weather conditions, merchant order priorities;

[0036] Reward function:

[0037]

[0038] Where α = 0.5, β = 0.3, γ = 0.2, S_safety is the safety factor;

[0039] The error rate of estimated arrival time ETA is controlled within ±3 minutes.

[0040] As a preferred solution of the present invention, the blockchain traceability module supports cross-chain interaction. When it detects that the temperature of the packaging box exceeds 4°C, it automatically triggers the synchronization of alliance chain data with the market supervision department to generate an electronic evidence chain including the following information:

[0041] Sensor raw data;

[0042] Delivery route trajectory;

[0043] Operator biometrics.

[0044] As a preferred solution of the present invention, the energy consumption management module of the central server simulates the carbon emissions of different distribution strategies through digital twin technology, generates a carbon neutrality report and optimizes the energy allocation plan to reduce carbon emissions per unit distribution mileage.

[0045] As a preferred solution of the present invention, the modular packaging box is made of PLA / PBAT biodegradable material, with a built-in RFID tag, which supports passive reading and writing in an environment of -20℃ to 8℃.

[0046] The present invention has the following beneficial effects:

[0047] 1. Full-dimensional environmental monitoring and risk warning: Through the collaboration of multi-parameter sensors such as temperature, collision, and gas (ethylene concentration, odor), three-dimensional perception of the transportation environment is achieved, breaking through the limitations of traditional cold chain single temperature monitoring, and identifying the risks of fresh food spoilage (such as abnormal ethylene concentration) and physical damage (such as drop impact) in advance, establishing a dual warning mechanism of "monitoring-prediction-intervention".

[0048] 2. Improved security and reliability:

[0049] Quantum encryption technology is combined with the electromagnetic locking system to prevent illegal opening and data tampering, solving the security vulnerabilities of traditional mechanical locks that are easily damaged and electronic locks that are susceptible to side channel attacks.

[0050] The redundant perception solution of lidar, millimeter-wave radar and visual cameras enhances the reliability of autonomous driving in complex road conditions (rain, snow, and night), breaking away from the limitations of single GPS navigation.

[0051] 4. Intelligent and efficient operations:

[0052] The dynamic adaptive buffer structure (shape memory alloy + aerogel) and intelligent partition layout (reinforcement learning algorithm) reduce physical damage to fresh produce and waste of transportation space, and improve loading efficiency and cargo protection capabilities.

[0053] Reinforcement learning path planning and multi-sensor fusion autonomous driving, combined with real-time traffic and order priorities, optimize delivery routes, shorten delivery time and reduce energy consumption.

[0054] 5. Data credibility and full-process traceability: Blockchain technology enables tamper-proof data storage throughout the entire process from production to consumption, supports cross-chain interaction and data synchronization with regulatory authorities, resolves the trust crisis of traditional centralized databases, and improves supply chain transparency.

[0055] 6. Environmental protection and sustainable development: PLA / PBAT biodegradable materials are used for packaging, combined with digital twin technology to optimize energy consumption distribution, forming a "green packaging + low-carbon transportation" closed loop, which complies with international plastic ban policy requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A block diagram of the composition of an unattended fresh food e-commerce cold chain system for merchants provided by an embodiment of the present invention;

[0057] Figure 2A temperature monitoring accuracy curve diagram for an unattended fresh food e-commerce cold chain system for merchants provided by an embodiment of the present invention;

[0058] Figure 3 This is a graph showing ethylene concentration detection and early warning for an unattended fresh food e-commerce cold chain system for merchants, as provided by an embodiment of the present invention;

[0059] Figure 4 This is a delivery time optimization curve diagram for an unattended fresh food e-commerce cold chain system for merchants provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The technical solution of the present invention will be further described below with reference to the accompanying drawings and through specific implementation methods.

[0061] Among them, the drawings are only used for illustrative purposes and represent only schematic diagrams rather than actual pictures, and should not be understood as limiting this patent; in order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0062] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "inside", "outside" and the like indicate an orientation or position relationship based on the orientation or position relationship shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0063] In the description of the present invention, unless otherwise expressly specified or limited, when the term "connection" or the like appears to indicate a connection relationship between components, such term should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be internal communication between two components or an interaction between two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood in specific circumstances.

[0064] This embodiment provides an unmanned fresh food e-commerce cold chain system for merchants, such as Figures 1-4 As shown, including: wherein, Figure 1 This is the block diagram of the system; Figure 2This is a temperature monitoring accuracy curve for this system; it shows the monitoring accuracy of the sensor at different temperatures, with the horizontal axis being the temperature range (-20°C to 20°C) and the vertical axis being the monitoring accuracy (±0.1°C). Figure 3 This is the ethylene concentration detection and warning curve of this system; the figure shows the warning time of the system under different ethylene concentrations, with the horizontal axis being ethylene concentration (0-500ppb) and the vertical axis being warning time (hours); Figure 4 This is a delivery time optimization curve for this system; it shows the ETA error rate of the system at different delivery distances, with the horizontal axis representing the delivery distance (0-100 kilometers) and the vertical axis representing the ETA error rate (±3 minutes). This unmanned fresh food e-commerce cold chain system for merchants includes a production system, a distribution center system, a delivery system, and a central server, including:

[0065] The production system is configured to standardize fresh produce through automatic sorting equipment and package the sorted produce into modular packaging boxes. The modular packaging boxes are equipped with a first sensor group, including: a three-axis MEMS accelerometer (range ±16g, resolution 0.01g), a metal oxide semiconductor gas sensor array (detection ethylene concentration range 0-500ppb), a fiber optic odor sensor (sensitivity ≤1ppm), and a low-power Bluetooth (BLE5.3) communication module;

[0066] The distribution center system includes a central distribution refrigerated container configured to receive and secure multiple modular packaging boxes. The refrigerated container is equipped with: a second sensor group, a multi-mode wireless communication module (supporting 4G / 5GNR+LoRa hybrid networking, with a dynamic switching threshold of ≤30dB), and an electromagnetic locking system (response time ≤50ms). The second sensor group includes a quartz crystal temperature sensor (accuracy of ±0.1°C), a fiber optic strain sensor (strain detection accuracy of ±5με), and a quantum encryption lock (using the BB84 protocol to resist side-channel attacks).

[0067] The distribution system includes unmanned cold chain vehicles loaded with central distribution refrigerated containers. These vehicles are equipped with: a reinforcement learning path planning module (based on the PPO algorithm, with a reward function integrating traffic congestion index and order timeliness) and a multi-sensor fusion autonomous driving unit (including a redundant solution of lidar, millimeter-wave radar, and visual cameras);

[0068] The central server integrates multi-source heterogeneous data through a federated learning framework, uses a spatiotemporal graph convolutional network (ST-GCN) to generate dynamic delivery routes, and triggers a three-level early warning mechanism based on a damage prediction model to reduce the fresh food loss rate.

[0069] The above technical solution achieves full-dimensional perception of the transportation environment through the coordinated monitoring of multiple parameter sensors (temperature, collision, and gas), solving the safety blind spots caused by traditional cold chain monitoring of only a single parameter (such as temperature). The combination of odor sensors and collision sensors can identify fresh produce spoilage risks (such as abnormal ethylene concentrations) and physical damage (such as drop impact) in advance, forming a dual early warning mechanism.

[0070] Quantum encryption technology prevents unauthorized opening and data tampering, addressing the vulnerability of traditional mechanical locks to damage and electronic locks to side-channel attacks. For the first time in logistics scenarios, quantum key distribution (QKD) is combined with physical locks to provide dual protection against theft and data security.

[0071] Redundant sensing solutions using LiDAR, millimeter-wave radar, and visual cameras enhance vehicle control reliability in complex road conditions (such as rainy, snowy, and nighttime delivery). This overcomes the limitations of traditional cold chain vehicles relying on single GPS navigation and enables autonomous delivery capabilities in all weather and all scenarios.

[0072] Distributed ledger technology ensures that data cannot be tampered with, resolving the trust crisis of traditional centralized databases that are vulnerable to attacks or internal tampering. By combining blockchain with cold chain logistics, it achieves transparent traceability of the entire process from production to consumption, enhancing trust between merchants and consumers.

[0073] The damage prediction model equation is as follows:

[0074]

[0075] in:

[0076] P damage : Probability of fresh product damage (0-1);

[0077] σ: Sigmoid activation function;

[0078] α k ,β i : Learnable weight parameters (dynamically updated through federated learning);

[0079] W k : spatiotemporal feature transformation matrix (generated by ST-GCN);

[0080] F i : the i-th eigenvector (including temperature change rate, vibration spectrum energy, transportation time, etc.);

[0081] γ: attention mechanism weight;

[0082] Attention(S t ,S p ): spatiotemporal attention function, calculated as:

[0083]

[0084] Q, K, V: query, key, value matrices (generated by GRU encoder);

[0085] d k : Dimension scaling factor.

[0086] Parameter description and technical effects:

[0087] Spatiotemporal feature transformation matrix W k

[0088] Definition: Extract spatiotemporal correlation features from multi-source data (temperature, humidity, vibration, path) through ST-GCN.

[0089] Technical effect: Breaking through the limitations of traditional models that rely solely on static features, it dynamically captures the interactive impact of delivery paths and the environment.

[0090] Example: During a delivery, the model identified the correlation between nighttime temperature fluctuations and the vibration spectrum on a certain road section, triggering an early warning.

[0091] Attention mechanism parameter γ

[0092] Definition: Adaptive weight parameter that aggregates the importance of data from each node through federated learning.

[0093] Technical effect: Dynamically adjust the contribution of characteristics such as temperature, vibration, and time to improve the model's sensitivity to key damage factors.

[0094] Example: During high temperature periods, the model automatically increases the weight of the temperature change rate (γ increases) to reduce the impact of transportation time.

[0095] Nonlinear activation function combination

[0096] Design: Tanh function constrains the feature range, and Sigmoid function outputs the probability value.

[0097] Advantages: Compared with traditional logistic regression, it introduces spatiotemporal feature transformation and nonlinear interaction to avoid linear assumption deviation.

[0098] Example scenario: A merchant transports fresh food packages from a warehouse to a community store, and the following data is recorded throughout the entire process:

[0099] Temperature change rate: ΔT = 0.8°C / h (exceeds the threshold by 0.5°C);

[0100] Vibration spectrum energy: E v =120m / s 2 (corresponding to an acceleration of 3.5g);

[0101] Transportation time: t=4.2h (3h longer than the recommended time).

[0102] Model calculation:

[0103] Input feature vector F = [ΔT,E v ,t];

[0104] ST-GCN extracts spatiotemporal features W k ·F;

[0105] The GRU encoder generates temporal dependencies S t , the path graph generates spatial dependence S p ;

[0106] Attention mechanism calculation Attention(S t ,S p )=0.78;

[0107] Output damage probability P damage =0.89, triggering the second-level warning (automatically allocate spare packaging boxes).

[0108] Working Principle Process

[0109] 1. Data collection: Obtain real-time data from sensor networks, GPS tracks, and environmental monitoring equipment;

[0110] 2. Feature preprocessing: normalization, missing value filling (using federated learning shared statistics);

[0111] 3. Spatiotemporal feature extraction: ST-GCN models path-environment associations, and GRU encodes temporal dependencies;

[0112] 4. Attention weighting: Dynamically adjust feature importance to generate a comprehensive injury risk score;

[0113] 5. Early warning decision: Trigger a graded response (adjust path / activate backup resources) based on probability thresholds.

[0114] Specifically, in this embodiment, the buffer structure of the modular packaging box includes: a shape memory alloy skeleton (Ni-Ti alloy, phase transition temperature 32°C) and an aerogel filling layer (silicon dioxide based, density ≤ 200kg / m 3 );

[0115] When the triaxial MEMS accelerometer detects a transient impact >5g, it expands within 200ms to form a honeycomb shock-absorbing structure through the shape memory effect to improve shock absorption efficiency.

[0116] By adopting the above technical solution and through the composite design of shape memory alloy and aerogel, a honeycomb shock-absorbing structure is automatically expanded when an impact is detected, thereby reducing the physical damage of fresh products during transportation, breaking through the limitations of traditional passive cushioning materials (such as foam), and realizing dynamic adaptive shock absorption.

[0117] Specifically, in this embodiment, the odor sensor adopts a dynamic sniffing algorithm and uses the XGBoost model to identify excessive ethylene concentrations, and warns of the risk of fruit and vegetable corruption 24 hours in advance to improve prediction accuracy.

[0118] By adopting the above technical solution, volatile organic compounds (VOCs) in the car can be monitored in real time, and the progress of fresh food spoilage can be predicted through changes in gas concentration, providing merchants with a basis for dynamic inventory management. Gas detection can be upgraded from environmental monitoring to a quality prediction tool, and cold chain management can be promoted from "after-the-fact remediation" to "pre-emptive intervention."

[0119] Specifically, in this embodiment, the adjustable partition plate of the central distribution refrigerated box is configured as follows:

[0120] A photoelectric sensor is installed at the end of the slide rail (detection accuracy ±1mm);

[0121] Dynamically optimize the packaging box layout based on reinforcement learning algorithms to improve space utilization.

[0122] By adopting the above technical solution, the cooperation between the slide rail and the photoelectric sensor can automatically optimize the space layout according to the size of the packaging box, reduce the transportation gap, increase the single delivery capacity, solve the space waste problem caused by traditional fixed partitions, and realize dynamic loading optimization.

[0123] Specifically, in this embodiment, the automatic driving unit of the unmanned cold chain vehicle includes:

[0124] Redundant braking system (hydraulic + electromagnetic dual braking, braking distance ≤ 3m / 100km / h);

[0125] Multimodal sensor data fusion module (Kalman filter algorithm, positioning accuracy ±10cm).

[0126] The above technical solution, with the dual protection of hydraulic and electromagnetic braking, provides fast and stable braking capabilities in emergency situations, reduces the risk of safety accidents caused by brake failure, introduces multi-braking mode redundancy design in cold chain distribution scenarios, and improves vehicle safety.

[0127] Specifically, in this embodiment, the three-level warning mechanism includes:

[0128] Level 1 warning: When the temperature fluctuation exceeds ±0.5°C, a suggestion to adjust the receiving time will be pushed to the merchant terminal;

[0129] Level 2 warning: When the collision intensity is detected to be greater than 3g or the ethylene concentration is greater than 100ppb, the spare sub-packaging box deployment process is initiated (response time ≤ 30 seconds);

[0130] Level 3 warning: When encountering illegal unlocking, the following will be executed simultaneously:

[0131] Release acidic electrolyte into the dispensing box (pH value drops below 2.5);

[0132] Adjust the direction of the air outlet of the cold chain vehicle's air conditioner to form a local low-temperature barrier (temperature gradient ≥ 5℃ / m).

[0133] Using the above technical solution, the first-level warning: temperature fluctuations prompt merchants to adjust their receiving plans to avoid receiving spoiled goods; the second-level warning: when collisions or gas abnormalities occur, spare packaging boxes are activated to ensure delivery continuity; the third-level warning: when illegal unlocking occurs, the security system is linked to prevent theft and preserve evidence; a "monitoring-warning-response" closed loop is built to achieve a leap from passive monitoring to active prevention and control.

[0134] Specifically, in this embodiment, the reinforcement learning algorithm of the path planning module includes:

[0135] State space definition: traffic flow, weather conditions, merchant order priority (weight coefficient 0.6-0.8) Reward function:

[0136]

[0137] Where α = 0.5, β = 0.3, γ = 0.2, S_safety is the safety factor;

[0138] The error rate of estimated time of arrival (ETA) is controlled within ±3 minutes.

[0139] By adopting the above technical solution, the optimal route is generated based on real-time traffic flow and order priority, which reduces delivery delays and energy waste, breaks through the traditional static route planning model, and improves delivery efficiency and resource utilization.

[0140] Specifically, in this embodiment, the blockchain traceability module supports cross-chain interaction. When it detects that the temperature of the packaging box exceeds 4°C, it automatically triggers data synchronization with the consortium chain of the market supervision department, generating an electronic evidence chain including the following information:

[0141] Sensor raw data (timestamp resolution ≤ 1ms);

[0142] Delivery route trajectory (GPS coordinate accuracy ±0.1m);

[0143] Operator biometrics (fingerprint / iris recognition records).

[0144] By adopting the above technical solution, when an abnormality is detected (such as temperature exceeding the standard), the data link will be automatically synchronized with the regulatory department to generate legally effective electronic evidence, realize the direct connection between cold chain data and the judicial system, and accelerate the quality dispute resolution process.

[0145] Specifically, in this embodiment, the energy consumption management module of the central server simulates the carbon emissions of different distribution strategies through digital twin technology, generates a carbon neutrality report and optimizes the energy allocation plan to reduce carbon emissions per unit distribution mileage.

[0146] By adopting the above technical solutions, the energy consumption of different distribution strategies can be simulated through digital twins, carbon footprint reports can be generated and the refrigeration mode of the cold chain can be optimized, so as to incorporate environmental protection indicators into the core of logistics decision-making and promote the sustainable development of the cold chain industry.

[0147] Specifically, in this embodiment, the modular packaging box is made of PLA / PBAT biodegradable material, has a built-in RFID tag (compliant with the EPCGen2 standard), and supports passive reading and writing in an environment of -20°C to 8°C (reading distance ≥ 5m).

[0148] Adopting the above technical solution, PLA / PBAT composite materials are used to make modular packaging boxes, which meet the requirements of international plastic ban policies and reduce the long-term pollution of traditional plastic packaging to the environment; built-in RFID tags that comply with the EPCGen2 standard support passive reading and writing (no external power supply required), and can work stably in a cold chain environment of -20℃ to 8℃, avoiding the endurance focus of traditional battery-powered tags, with a reading distance of ≥5 meters, which is suitable for the rapid inventory needs of batch packaging boxes of cold chain vehicles, reducing the workload of manual scanning and improving logistics efficiency; the EPCGen2 standard supports cross-regional and cross-system data interaction, ensuring that the packaging boxes can be recognized by reading and writing equipment in different countries during cross-border transportation, meeting the requirements of international logistics standardization, and combining the regional Blockchain technology enables the data of the entire life cycle of packaging boxes, from production and sorting to distribution, to be uploaded to the chain, enhancing supply chain transparency and merchant trust; the introduction of PBAT reduces the use of PLA (PBAT costs 30% less than PLA), and at the same time improves the mechanical strength of the packaging boxes (such as puncture resistance) through the composite process, reducing the transportation damage rate. PLA / PBAT composite materials still maintain flexibility at low temperatures (-20°C), avoiding the problem of brittle cracking of traditional plastics in the cold chain; degradable materials reduce methane emissions from plastic waste landfills, and combined with the energy consumption management module of the central server, form a low-carbon closed loop from packaging to transportation, complying with domestic and international environmental regulations, and avoiding fines or market access restrictions caused by the use of non-environmentally friendly materials.

[0149] In summary, the unmanned fresh food e-commerce cold chain system for merchants provided in this embodiment has the following advantages:

[0150] 1. Full-dimensional environmental monitoring and risk warning: Through the collaboration of multi-parameter sensors such as temperature, collision, and gas (ethylene concentration, odor), three-dimensional perception of the transportation environment is achieved, breaking through the limitations of traditional cold chain single temperature monitoring, and identifying the risks of fresh food spoilage (such as abnormal ethylene concentration) and physical damage (such as drop impact) in advance, establishing a dual warning mechanism of "monitoring-prediction-intervention".

[0151] 2. Improved security and reliability:

[0152] Quantum encryption technology is combined with the electromagnetic locking system to prevent illegal opening and data tampering, solving the security vulnerabilities of traditional mechanical locks that are easily damaged and electronic locks that are susceptible to side channel attacks.

[0153] The redundant perception solution of lidar, millimeter-wave radar and visual cameras enhances the reliability of autonomous driving in complex road conditions (rain, snow, and night), breaking away from the limitations of single GPS navigation.

[0154] 4. Intelligent and efficient operations:

[0155] The dynamic adaptive buffer structure (shape memory alloy + aerogel) and intelligent partition layout (reinforcement learning algorithm) reduce physical damage to fresh produce and waste of transportation space, and improve loading efficiency and cargo protection capabilities.

[0156] Reinforcement learning path planning and multi-sensor fusion autonomous driving, combined with real-time traffic and order priorities, optimize delivery routes, shorten delivery time and reduce energy consumption.

[0157] 5. Data credibility and full-process traceability: Blockchain technology enables tamper-proof data storage throughout the entire process from production to consumption, supports cross-chain interaction and data synchronization with regulatory authorities, resolves the trust crisis of traditional centralized databases, and improves supply chain transparency.

[0158] 6. Environmental protection and sustainable development: PLA / PBAT biodegradable materials are used for packaging, combined with digital twin technology to optimize energy consumption distribution, forming a "green packaging + low-carbon transportation" closed loop, which complies with international plastic ban policy requirements.

[0159] The workflow is as follows:

[0160] 1. Production system

[0161] Standardized processing: Fresh products are screened, graded, and packaged into modular packaging boxes through automatic sorting equipment.

[0162] Intelligent packaging: The packaging box is equipped with a three-axis acceleration sensor, gas sensor, odor sensor and BLE communication module to collect data such as collision, ethylene concentration, and volatile organic compounds (VOCs) in real time; the buffer structure uses shape memory alloy and aerogel, and automatically deploys a honeycomb shock-absorbing structure when impact occurs.

[0163] 2. Distribution center system

[0164] Goods Receiving: The central distribution refrigerated container receives the modular packaging boxes and is fixed with an electromagnetic locking system. Built-in temperature sensors (quartz crystal) and strain sensors (optical fiber) monitor the environment and physical status of the refrigerated container.

[0165] Space optimization: Adjustable dividers work with slides and photoelectric sensors to dynamically adjust the layout based on reinforcement learning algorithms to improve space utilization.

[0166] Communication and security: A multi-mode wireless communication module (4G / 5G+LoRa) enables data upload and command reception, and a quantum encryption lock ensures the physical and data security of the reefer.

[0167] 3. Distribution system

[0168] Autonomous driving: Cold chain vehicles use a combination of lidar, millimeter-wave radar, and visual cameras to perceive road conditions, and use the PPO reinforcement learning algorithm to plan routes (integrating traffic congestion and order priorities) to achieve all-weather autonomous delivery.

[0169] Safety response: Redundant braking systems (hydraulic + electromagnetic) ensure emergency braking; a three-level warning mechanism triggers different responses based on sensor data (such as temperature fluctuation warning, dispatching spare packaging boxes in case of collision / gas abnormality, and releasing electrolyte and activating low-temperature barriers in case of illegal unlocking).

[0170] 4. Central server link

[0171] Data integration: Aggregate multi-source heterogeneous data (temperature, location, sensor status, etc.) from production, collection, and distribution systems through a federated learning framework.

[0172] Intelligent decision-making: The spatiotemporal graph convolutional network (ST-GCN) generates dynamic delivery routes; the damage prediction model combines with the XGBoost algorithm to identify abnormal ethylene concentrations and provide early warning of corruption risks.

[0173] Traceability and management: The blockchain module records full-process data (sensor raw data, delivery trajectory, operator biometrics), supports cross-chain interaction and regulatory synchronization; digital twin technology simulates carbon emissions, optimizes energy distribution plans, and generates carbon neutrality reports.

[0174] 5.Terminal interaction

[0175] Merchant side: Receive first-level warning (temperature fluctuations suggest adjusting the receiving time) and second-level warning (activate backup packaging boxes), and verify the quality of fresh produce through blockchain traceability data.

[0176] Supervision side: Abnormal events (such as excessive temperature, illegal unlocking) trigger automatic data synchronization, generate a legally binding electronic evidence chain, and accelerate the handling of quality disputes.

[0177] The entire process forms a closed loop through "real-time monitoring by sensors - intelligent decision-making by central servers - coordinated execution of various systems", realizing intelligent, safe and efficient management of the entire chain from production to distribution.

[0178] The above are only preferred embodiments of the present invention and do not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. An unmanned fresh food e-commerce cold chain system for merchants, characterized by: include: A production system configured to standardize fresh produce using automatic sorting equipment and package the sorted produce into modular packaging boxes, wherein the modular packaging boxes have a built-in first sensor group including: a three-axis MEMS accelerometer, a metal oxide semiconductor gas sensor array, a fiber optic odor sensor, and a communication module; The distribution center system includes a central distribution refrigerated container configured to receive and secure a plurality of modular packaging boxes, wherein the refrigerated container is equipped with: a second sensor group, a multi-mode wireless communication module, and an electromagnetic locking system, wherein the second sensor group includes a quartz crystal temperature sensor, an optical fiber strain sensor, and a quantum encryption lock; The distribution system includes an unmanned cold chain vehicle loaded with central distribution refrigerated containers, and the cold chain vehicle is equipped with a reinforcement learning path planning module and a multi-sensor fusion autonomous driving unit; The central server integrates multi-source heterogeneous data through a federated learning framework, uses the spatiotemporal graph convolutional network (ST-GCN) to generate dynamic delivery routes, and triggers a three-level early warning mechanism based on a damage prediction model to reduce the fresh food loss rate.

2. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The buffer structure of the modular packaging box includes: Shape memory alloy skeleton; Aerogel filling layer; When the triaxial MEMS accelerometer detects a transient impact >5g, it expands within 200ms to form a honeycomb shock-absorbing structure through the shape memory effect to improve shock absorption efficiency.

3. The unattended fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The odor sensor uses a dynamic sniffing algorithm and an XGBoost model to identify excessive ethylene concentrations, providing a 24-hour advance warning of the risk of fruit and vegetable spoilage to improve prediction accuracy.

4. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The adjustable partition plate of the central distribution refrigerated box is configured as follows: A photoelectric sensor is installed at the end of the slide rail; Dynamically optimize the packaging box layout based on reinforcement learning algorithms to improve space utilization.

5. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The automatic driving unit of the unmanned cold chain vehicle includes: Redundant braking system; Multimodal sensor data fusion module.

6. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The three-level early warning mechanism includes: Level 1 warning: When the temperature fluctuation exceeds ±0.5°C, a suggestion to adjust the receiving time will be pushed to the merchant terminal; Level 2 warning: When the collision intensity is detected to be greater than 3g or the ethylene concentration is greater than 100ppb, the spare sub-packaging box allocation process is started; Level 3 warning: When encountering illegal unlocking, the following will be executed simultaneously: Release acid electrolyte into the dispensing box; Adjust the direction of the air outlet of the cold chain vehicle's air conditioner to form a local low-temperature barrier.

7. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The reinforcement learning algorithm of the path planning module includes: State space definition: traffic flow, weather conditions, merchant order priorities; Reward function: Where α = 0.5, β = 0.3, γ = 0.2, S_safety is the safety factor; The error rate of estimated arrival time ETA is controlled within ±3 minutes.

8. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The blockchain traceability module supports cross-chain interaction. When it detects that the temperature of the packaging box exceeds 4°C, it automatically triggers data synchronization with the alliance chain of the market supervision department to generate an electronic evidence chain including the following information: Sensor raw data; Delivery route trajectory; Operator biometrics.

9. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The energy consumption management module of the central server simulates the carbon emissions of different distribution strategies through digital twin technology, generates carbon neutrality reports and optimizes energy allocation plans to reduce carbon emissions per unit distribution mileage.

10. The unmanned fresh food e-commerce cold chain system for merchants according to claim 1 is characterized in that: The modular packaging box is made of PLA / PBAT biodegradable materials, has a built-in RFID tag, and supports passive reading and writing in an environment of -20°C to 8°C.

Images